Skip to content

Add scheduling mechanism and new workload #1025

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
helloyongyang merged 11 commits into from
Apr 24, 2026
Merged

Add scheduling mechanism and new workload #1025

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
3b9db5a
add elastic scheduling and make the system more stable
zhtshr Mar 31, 2026
40557eb
add test
zhtshr Apr 14, 2026
0cdea75
fix scheduler bugs
zhtshr Apr 20, 2026
9972f81
Merge branch 'zht_dev'
zhtshr Apr 20, 2026
b24f32b
fix merge bugs
zhtshr Apr 20, 2026
a7bf507
test multi machine
zhtshr Apr 20, 2026
fe25221
multi machine
zhtshr Apr 23, 2026
5a90f98
enable multi machine test
zhtshr Apr 24, 2026
04f98e7
run precommit
zhtshr Apr 24, 2026
5849577
Merge branch 'main' into main
zhtshr Apr 24, 2026
ee4ed2d
run precommit
zhtshr Apr 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
158 changes: 158 additions & 0 deletions configs/disagg/multi_node/wan22_i2v_distill_controller.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,158 @@
{
"infer_steps": 4,
"in_dim": 36,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"num_heads": 40,
"num_layers": 40,
"out_dim": 16,
"eps": 1e-06,
"model_type": "i2v",
"target_video_length": 81,
"text_len": 512,
"target_height": 480,
"target_width": 832,
"self_attn_1_type": "sage_attn2",
"cross_attn_1_type": "sage_attn2",
"cross_attn_2_type": "sage_attn2",
"sample_guide_scale": [
3.5,
3.5
],
"sample_shift": 5.0,
"enable_cfg": false,
"cpu_offload": true,
"offload_granularity": "block",
"rdma_buffer_slot_size": 8192,
"t5_cpu_offload": false,
"vae_cpu_offload": false,
"fps": 16,
"use_image_encoder": false,
"boundary_step_index": 2,
"denoising_step_list": [
1000,
750,
500,
250
],
"dit_quantized": true,
"dit_quant_scheme": "int8-q8f",
"high_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"high_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"image_path": "/root/zht/LightX2V/assets/inputs/imgs/img_0.jpg",
"disagg_mode": "controller",
"disagg_config": {
"bootstrap_addr": "192.168.0.166",
"bootstrap_room": 0,
"ranks": 8,
"encoder_engine_rank": 0,
"transformer_engine_rank": 1,
"decoder_engine_rank": 2,
"protocol": "rdma",
"local_hostname": "192.168.0.166",
"metadata_server": "P2PHANDSHAKE",
"service_env": {
"RDMA_IFACE": "erdma_0"
},
"remote_workdir": "/root/zht/LightX2V",
"remote_python_executable": "python",
"remote_activate_cmd": "source /root/miniconda3/etc/profile.d/conda.sh && conda activate lightx2v && export LD_LIBRARY_PATH=/root/miniconda3/envs/lightx2v/lib:${LD_LIBRARY_PATH:-}",
"remote_log_dir": "/root/zht/LightX2V/save_results",
"use_remote_proxy": true,
"remote_proxy_req_base_port": 28000,
"ssh_user": "root",
"ssh_options": [
"-i",
"/root/.ssh/id_ed25519_zht",
"-o",
"BatchMode=yes",
"-o",
"StrictHostKeyChecking=no"
],
"static_instance_slots": [
{
"instance_type": "encoder",
"host": "192.168.0.139",
"engine_rank": 0,
"cuda_device": 0,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.139"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 1,
"cuda_device": 0,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 2,
"cuda_device": 1,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 3,
"cuda_device": 2,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 4,
"cuda_device": 3,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 5,
"cuda_device": 4,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "transformer",
"host": "192.168.0.166",
"engine_rank": 6,
"cuda_device": 5,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.166"
}
},
{
"instance_type": "decoder",
"host": "192.168.0.139",
"engine_rank": 7,
"cuda_device": 1,
"env": {
"MOONCAKE_DEVICE_NAME": "eth0",
"MOONCAKE_LOCAL_HOSTNAME": "192.168.0.139"
}
}
]
}
}
58 changes: 58 additions & 0 deletions configs/disagg/multi_node/wan22_i2v_distill_decoder.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
{
"infer_steps": 4,
"in_dim": 36,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"num_heads": 40,
"num_layers": 40,
"out_dim": 16,
"eps": 1e-06,
"model_type": "i2v",
"target_video_length": 81,
"text_len": 512,
"target_height": 480,
"target_width": 832,
"self_attn_1_type": "sage_attn2",
"cross_attn_1_type": "sage_attn2",
"cross_attn_2_type": "sage_attn2",
"sample_guide_scale": [
3.5,
3.5
],
"sample_shift": 5.0,
"enable_cfg": false,
"cpu_offload": true,
"offload_granularity": "block",
"rdma_buffer_slot_size": 8192,
"t5_cpu_offload": false,
"vae_cpu_offload": false,
"fps": 16,
"use_image_encoder": false,
"boundary_step_index": 2,
"denoising_step_list": [
1000,
750,
500,
250
],
"dit_quantized": true,
"dit_quant_scheme": "int8-q8f",
"high_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"high_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"image_path": "/root/zht/LightX2V/assets/inputs/imgs/img_0.jpg",
"disagg_mode": "decoder",
"disagg_config": {
"bootstrap_addr": "192.168.0.166",
"bootstrap_room": 0,
"ranks": 8,
"encoder_engine_rank": 0,
"transformer_engine_rank": 1,
"decoder_engine_rank": 2,
"protocol": "rdma",
"local_hostname": "192.168.0.139",
"metadata_server": "P2PHANDSHAKE"
}
}
58 changes: 58 additions & 0 deletions configs/disagg/multi_node/wan22_i2v_distill_encoder.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
{
"infer_steps": 4,
"in_dim": 36,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"num_heads": 40,
"num_layers": 40,
"out_dim": 16,
"eps": 1e-06,
"model_type": "i2v",
"target_video_length": 81,
"text_len": 512,
"target_height": 480,
"target_width": 832,
"self_attn_1_type": "sage_attn2",
"cross_attn_1_type": "sage_attn2",
"cross_attn_2_type": "sage_attn2",
"sample_guide_scale": [
3.5,
3.5
],
"sample_shift": 5.0,
"enable_cfg": false,
"cpu_offload": true,
"offload_granularity": "block",
"rdma_buffer_slot_size": 8192,
"t5_cpu_offload": false,
"vae_cpu_offload": false,
"fps": 16,
"use_image_encoder": false,
"boundary_step_index": 2,
"denoising_step_list": [
1000,
750,
500,
250
],
"dit_quantized": true,
"dit_quant_scheme": "int8-q8f",
"high_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"high_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"image_path": "/root/zht/LightX2V/assets/inputs/imgs/img_0.jpg",
"disagg_mode": "encoder",
"disagg_config": {
"bootstrap_addr": "192.168.0.166",
"bootstrap_room": 0,
"ranks": 8,
"encoder_engine_rank": 0,
"transformer_engine_rank": 1,
"decoder_engine_rank": 2,
"protocol": "rdma",
"local_hostname": "192.168.0.139",
"metadata_server": "P2PHANDSHAKE"
}
}
58 changes: 58 additions & 0 deletions configs/disagg/multi_node/wan22_i2v_distill_transformer.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
{
"infer_steps": 4,
"in_dim": 36,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"num_heads": 40,
"num_layers": 40,
"out_dim": 16,
"eps": 1e-06,
"model_type": "i2v",
"target_video_length": 81,
"text_len": 512,
"target_height": 480,
"target_width": 832,
"self_attn_1_type": "sage_attn2",
"cross_attn_1_type": "sage_attn2",
"cross_attn_2_type": "sage_attn2",
"sample_guide_scale": [
3.5,
3.5
],
"sample_shift": 5.0,
"enable_cfg": false,
"cpu_offload": true,
"offload_granularity": "block",
"rdma_buffer_slot_size": 8192,
"t5_cpu_offload": false,
"vae_cpu_offload": false,
"fps": 16,
"use_image_encoder": false,
"boundary_step_index": 2,
"denoising_step_list": [
1000,
750,
500,
250
],
"dit_quantized": true,
"dit_quant_scheme": "int8-q8f",
"high_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_quantized_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"high_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_high_noise_int8_lightx2v_4step.safetensors",
"low_noise_original_ckpt": "/root/zht/LightX2V/models/lightx2v/Wan2.2-Distill-Models/wan2.2_i2v_A14b_low_noise_int8_lightx2v_4step.safetensors",
"image_path": "/root/zht/LightX2V/assets/inputs/imgs/img_0.jpg",
"disagg_mode": "transformer",
"disagg_config": {
"bootstrap_addr": "192.168.0.166",
"bootstrap_room": 0,
"ranks": 8,
"encoder_engine_rank": 0,
"transformer_engine_rank": 1,
"decoder_engine_rank": 2,
"protocol": "rdma",
"local_hostname": "192.168.0.166",
"metadata_server": "P2PHANDSHAKE"
}
}
0 ...igs/disagg/wan_t2v_disagg_controller.json → ...multi_node/wan_t2v_disagg_controller.json
View file
Open in desktop
File renamed without changes.
0 configs/disagg/wan_t2v_disagg_decoder.json → ...gg/multi_node/wan_t2v_disagg_decoder.json
View file
Open in desktop
File renamed without changes.
0 configs/disagg/wan_t2v_disagg_encoder.json → ...gg/multi_node/wan_t2v_disagg_encoder.json
View file
Open in desktop
File renamed without changes.
0 ...gs/disagg/wan_t2v_disagg_transformer.json → ...ulti_node/wan_t2v_disagg_transformer.json
View file
Open in desktop
File renamed without changes.
Loading
Loading